Navigazione – Mappa del sito

HomeNumeri43ontologia dei coloriStuck in the middle: Colors betwe...

ontologia dei colori

Stuck in the middle: Colors between the subjective and the objective

Eyja M. Brynjarsdóttir
p. 47-65


I argue that there are good reasons for thinking of colors as both subjective and objective. I propose a spectrum ranging from the entirely subjective to the entirely objective, on which colors belonging somewhere between the two ends. I then argue that these findings about colors can be applied to other sensory properties as well because the reasons for placing colors where they belong on the spectrum between the subjective and the objective hold for all sensory properties.

Torna su

Testo integrale

  • 1  We can suppose that the virus causes genetic changes as well so that future generations of humans (...)
  • 2  See Blackburn 1985: 14 for someone whose intuition seems to be that the color of grass would have (...)
  • 3  An example of such a view can be found in Byrne and Hilbert 1997.

1Suppose all humans were afflicted with a highly contagious virus causing permanent changes in their visual system1. The virus would cause a spectrum inversion; grass now looked to us to have the color that ripe tomatoes looked to have before the plague. Would grass still be green after this change or would it have become red even though its intrinsic properties had not changed? Answers to this question, often based on intuition, seem to vary2. Those who think colors would stay the same through this kind of change in our color vision can be called color objectivists. They think of colors as mind-independent; as properties things have independently of the way we experience them. Among color objectivists are those who endorse some version of the view that colors are microphysical properties3 as well as those who subscribe to the view described below as primitivism. Those who think the colors of things would have changed because of the change in our visual system can be called color subjectivists. On their view, the colors of things depend on our experiences of them; colors are mind-dependent.

2The aim of this paper is to explain why both objectivists and subjectivists about color have some claim to being right. The reason is that our experiences of color play an important role in determining the colors of things, yet it seems plausible that colors are mind-independent to a significant degree. Before outlining this reason, I discuss two theories of color that I consider false yet illuminating: color primitivism and error theories. I then propose a spectrum ranging from the objective to the subjective, and claim that many properties fall somewhere in its middle. In the last third of the paper, I argue that these findings about color can be extrapolated to other sensory properties.


  • 4  Johnston 1992: 138.

3In his paper How to Speak of the Colors (Johnston 1992), Mark Johnston offers a list of what he calls our core beliefs about color. Those are beliefs we have about color and a property of which they are not true cannot be color, i.e. it must be some other property. Among those beliefs is one Johnston calls Revelation: «The intrinsic nature of canary yellow is fully revealed by a standard visual experience as of a canary yellow thing»4. The idea is that via our visual experiences of a color we get to know all there is to know about the nature of the color. Johnston claims that while we are tempted to believe Revelation, its inconsistency with some other core beliefs shows that we must abandon it. In order to save as many core beliefs about color as possible, Revelation must be sacrificed. Even though it may be tempting to believe that the full nature of color is directly revealed to us in visual experience, it just cannot be true. Johnston argues that despite this there is a grain of truth in Revelation and he proposes a qualified version of it: Visual experience does not reveal to us everything there is to know about the nature of color, but it still reveals to us important parts of its nature, such as similarity and difference relations between colors.

  • 5  Examples of color primitivists are Campbell 1993, McGinn 1996, Strawson G. 1989, Stroud 2000, and (...)
  • 6  Error theorists include Boghossian and Velleman 1989, Mackie 1976, and Hardin 1988.

4While Johnston thinks Revelation must be abandoned, the belief has its followers. Those who think that Revelation is the standard, common sense belief about color belong in two different camps. On the one hand, we have those who take Revelation as a true belief, i.e. who think not only that Revelation is the standard, everyday view, but also that we are right to believe it. According to such a view, colors really are simple, mind-independent properties that are exactly like our phenomenal experiences of them5. I will be calling that view color primitivism. On the other hand, we have those who think that we believe something like Revelation about color, but that the objects around us do not have any corresponding properties. I refer to this view as an error theory6.

5In his paper A Simple View of Colour, John Campbell claims that our experiences of colors show us colors just the way they are:

  • 7  Campbell 1993: 257.

Still, if we take the appearances at face value, we will take it that we are seeing the properties of objects in virtue of which they have the potential to produce experiences of colour. The perception reveals the whole character of the property to us7.

6Campbell argues that since this transparency thesis holds, colors can neither be microphysical properties nor dispositions to cause experiences:

  • 8  Campbell 1993: 258.

A simpler view of colours thus remains in play. On this view, redness, for example, is not a disposition to produce experiences in us. It is, rather, the ground of such a disposition. But that is not because redness is a microphysical property – the real nature of the property is, rather, transparent to us8.

  • 9  McGinn 1996: 538.

7According to Campbell, colors are simple properties that are exactly as presented by our color experiences. Another proponent of color primitivism, Colin McGinn, rejects the dispositional view of color for similar reasons: «we just do not see colors as dispositions to cause experiences»9, and finds any account of color unacceptable unless it claims that colors are exactly as we perceive them to be.

8From this we may gather that according to color primitivism, colors are intrinsic properties of the objects we see. And these intrinsic properties are exactly as we perceive them; i.e. they are just like the sensation we have when we see them, just the way they are phenomenologically presented to us. Our perceptions give us direct access to colors exactly the way they are. The phenomenology of color vision is to be taken at face value: as we attend to our experiences of colors, we can infer that exactly this is how the colors are.

9Many arguments have been made against color primitivism, such as that it makes it impossible for colors to be the causes of our color experiences, that it makes it very unclear what kind of properties colors should be, and that it is simply based on wrong assumptions about the way color vision works. But what I take to be the most serious problem with color primitivism is the idea of mind-independent, intrinsic properties to which we have direct perceptual access.

  • 10  Edwards 1998; 2003. Arguments against Edwards 1998, to which Edwards responds in his 2003, can be (...)
  • 11  Edwards 2003: 110.

10Jim Edwards has argued in a powerful way that primitivism is incompatible with semantic externalism10. Edwards asks us to imagine a planet he calls Z-land. In Z-land, things are just like they are on our planet except for one thing: the light in Z-land consists of Z-rays that have a color inverting effect. When someone in Z-land is looking at the grass, it looks just the way red things look here. The inhabitants of Z-land speak of the grass as green* and of poppy flowers as red*. If we assume that color primitivism holds, the sense of the term “green*” is that it applies to things that have the property we call redness (because the sense is determined by the Z-landers’ experiences of things such as grass), yet the reference of “green*” is the property we call greenness (remember that the grass in Z-land is just like here; it’s just that the Z-rays make it look different). This means that there is a mismatch between sense and reference and the truth conditions of “This is green*” cannot be determined as they differ depending on whether we rely on the sense or the reference of “green*”. This does not happen if primitivism is abandoned, as in that case the sense and reference can be defined so that they match. For instance, the sense of “green*” could then be «being disposed to look red to standard perceivers in Z-land» which does not conflict with the reference being the property of greenness11.

11The important point suggested by Edwards’s argument is that if something is to be mind-independent, there must be a way of separating it from how it is experienced. In order for the greenness of the grass to be intrinsic to the grass and independent of our (or the Z-landers’) perceptions of grass, it must be freed from the requirement of being just like those perceptions. If greenness is truly mind-independent, it must be possible for it to fail under some circumstances to produce any sensations of green. Something that is mind-independent can only contingently produce a certain kind of subjective experience; that is what makes it mind-independent. Defining greenness as mind-independent and at the same time by definition identical to the subjective experience it contingently produces does not work.

12One of Edwards’s points is that if the reference of a term is externally determined, it is determined independently of the subject’s experiences. Hence it cannot be defined in the same way as the internally determined sense of the term. The efficacy of this version of the argument rests on our wanting to retain semantic externalism. While that may well be a worthy cause, I will now outline a version of the argument that does not rely on semantic externalism:

  • 12  A related criticism can be found in Chalmers 2006: 67-68.

13Imagine Zoë the Z-lander, an avid color primitivist. When she looks at grass, her experience is phenomenally the same as that of her earthly counterpart when she looks at a red poppy. This means that Zoë believes that grass has an intrinsic property that matches exactly this experience of hers. But remember that the grass in Z-land is intrinsically just like the grass on Earth. Now, let us assume that color primitivism is true, and that the color of grass is green; just like the experiences that we Earthlings have when we look at grass. If this is the case, Zoë must be wrong. The grass in Z-land does not have an intrinsic property that is just like Zoë’s experience of it (on the other hand, poppies do, but Zoë does not know that). This seems to mean that an error theory about color is true in Z-land: the Z-landers are systematically mistaken about colors. But saying that the Z-landers are the ones who are mistaken rather than the Earthlings seems completely arbitrary. If greenness is mind-independent, then grass’ being green has nothing more to do with us Earthlings than with Z-landers, so why should we presume to be the lucky ones who have it right12? Perhaps the rays of our star, Sol, are those that invert the colors. Hence, we cannot even distinguish for sure between primitivism and an error theory.

14While the idea of mind-independent properties being identical with subjective experiences is problematic, primitivism about subjective properties seems more plausible. Suppose for a moment that colors were mind-dependent so that an object would be green if and only if someone perceived it as green, i.e. had phenomenally green responses to it. In that case, assuming that greenness is exactly like our sensations of green is less of a problem. Grass is green on Earth, at least during the day for most people, and poppies are green in Z-land for most people most of the time. If our perceptions cause the property to be instantiated, it is not a problem for the property to be exactly like our perceptions of it.

15I conclude from this that while primitivism might be true for some subjective properties, it cannot be true for mind-independent properties. Note that this is not about whether we can know all there is to know about the nature of the property in question. In some cases, we know everything about what constitutes an instantiation of a property because we have made and/or defined it ourselves, yet the property can be objectively instantiated. An example could be the property of being legal. Obviously, our knowing all about what makes something legal would not make the property of being legal exactly like the phenomenal content of our experiences of legality. There is no such phenomenal content and no such experience. Assuming that a property is exactly like the way we think of it does not amount to Revelation about it, and the view that we are correct to assume that does not amount to primitivism. We may judge or infer that something is legal, but there is no such thing as the way it is to perceive something legal. Revelation and primitivism are only applicable to properties that can be associated with a sensation.

Error Theories

16According to error theorists about color, we believe Revelation, that colors are mind-independent properties whose nature is revealed in visual perception, but we are mistaken in that belief. The objects around us do not have any mind-independent properties just like our sensations of color; therefore we are guilty of a systematic error. I have just argued that there are no such mind-independent properties as those described by Revelation. Does that amount to an error theory of color?

17The short answer is no. The long answer is that there are a couple of reasons to reject an error theory of color. Let me elaborate:

  • 13  Chalmers 2006: 50.
  • 14Ibidem.
  • 15  Chalmers 2006: 66-69.

18In his paper Perception and the Fall from Eden (Chalmers 2006), David Chalmers argues that our experiences have several different representational contents. One of them is their phenomenal content: «A phenomenal content of a perceptual experience is a representational content that is determined by the experience’s phenomenal character»13. Chalmers defines representational content as a condition of satisfaction of the experience in question, and phenomenal character as what it is like to have the experience14. If the objects we see around us had properties that were exactly like the phenomenal content of our color experiences, these properties would be what Chalmers calls perfect colors and our experiences of them would be perfectly veridical. However, things do not have perfect colors (if they did, color primitivism would be true) and our color experiences are not perfectly veridical15.

  • 16  Chalmers 2006: 72.

19Even though our color experiences are not perfectly veridical, they can be imperfectly veridical, says Chalmers. They are imperfectly veridical if the objects we see have properties that are good enough to serve as matches for the phenomenal content, even though they do not bear a perfect resemblance to it. Instead of a red apple having the property of perfect redness it could be the case that it has the property of imperfect redness. Chalmers accomplishes this by attributing two phenomenal contents to each experience: Edenic content, which can only be satisfied by the object’s having a perfectly resembling property; and Fregean content, a mode of presentation of the object and its properties determined by the Edenic content16. An experience of redness is veridical if and only if the object of the experience possesses a property that matches perfect redness, i.e. it normally causes phenomenally red experience. It is, however, not perfectly veridical unless the object has a property perfectly matching the Edenic content. Unless this second requirement is fulfilled, the experience is imperfectly veridical, which is really all we can hope for.

20This is what Chalmers calls the two stage view of phenomenal content. This two stage view describes one way an error theory about color can be avoided. Even though the objects around us do not have properties that are perfect resemblances of the way they are phenomenologically presented to us, or what Chalmers would call Edenic content, they can have properties that serve well enough as matches for those experiences. As long as our color terms successfully and consistently refer to properties that objects actually have, it does not seem necessary for those properties to perfectly resemble the phenomenal content of our color experiences.

21My main reason, however, for rejecting an error theory about color is that I do not think we actually believe Revelation about color. If we really believed that our color terms referred to objective properties perfectly resembling the phenomenal content of our color experiences, and then the objects around us did not have any such properties, there would be an error in play. Our beliefs about color would be systematically false. But why should we assume that we believe Revelation?

  • 17  Of course, not just any causal chain will do in this respect. The father could for instance cause (...)

22Imagine a little boy’s drawing of his father. The drawing does not look much like the father; it shows a squiggly stick figure. However, we accept the drawing as a representation of the father. Presumably, the reason is that we can account for an appropriate causal chain17 between the father and what we see in the picture, and that chain is sufficiently strong to make the squiggly stick figure on the sheet of paper a true representation of the father.

23If the boy believed that the drawing showed exactly the way his father really looked, there might be an error involved. But that would have to be a very literal belief. We would even accept claims such as “This is how I see him” as not involving an error. Similarly, if we literally believed that grass had a property that was exactly like the way greenness is phenomenologically presented to us, we would believe Revelation and be in error. But the belief that grass has a property that is presented to us in this particular way does not entail the belief that the property literally is exactly as we picture it. We can believe that the way greenness is phenomenologically presented to us is representative of a property in the grass without believing that the property of the grass really is exactly like that. And why should we believe the latter?

Color Concepts and Color Properties

  • 18  My view here is similar to that of Johnston’s mentioned above.

24While neither color primitivism nor an error theory of color is a viable option, both views express something important about our color concepts. Both views involve the idea that we generally believe Revelation about colors. While I doubt that we really do, I think we do believe that our color sensations show us something important about colors. Even though we do not strictly believe that the colors are exactly like our sensations of them, we do think we can make all sorts of judgments about them purely on the basis of those sensations18. One example is relations of similarity and difference between the colors: We take it for granted that orange is more similar to red than it is to blue simply because our sensation of orange is more similar to that of red than to that of blue. Another example is the belief that colors are visible; that we can tell by sight, at least most of the time under most conditions, which colors objects have. That is, we believe that our sensations of blue give us, at least usually, the information that the object we are looking at is blue.

25What matters here is that we do not have to believe Revelation about color in order to believe that there is an extent to which our sensations of color tell us exactly what the colors are and what they are like. We can believe that our color sensations reveal to us everything about the colors that matters to us for most of our everyday purposes. According to that belief, our color sensations reveal to us which colors are similar and which are different, which colors “go well” together and which do not, that when we mix blue paint with yellow we get green paint etc., and for those practical, mundane purposes it may be sufficient for me to point to something blue and say “look at that” in order to explain what it is for something to be blue. But this does not mean that we really believe that our color sensations actually reveal all there is to know about the nature of color. It no more does than the claim of an experienced author that she knows all about book writing commits her to the claim that she knows everything there is to know about all the things involved, such as the writer’s brain processes or the hardware of the computer used to write the books. Even though we may believe that color sensations reveal everything that matters about the colors in a certain context, it does not commit us to such a belief in all contexts, or the belief that colors are nothing above and beyond what color sensations reveal to us. And it does not commit us to the belief that colors actually are properties that perfectly resemble our color sensations.

26The fact that there is a context in which we believe that our color sensations reveal everything, or at least everything that matters, about the colors suggests that our color sensations play an important role in our color concepts. It is essential to our concept of blueness that it is of a property responsible for our sensations of blue. A property that is not behind those phenomenological experiences of ours has to be something other than blueness (this is similar to what Johnston and others have said about core beliefs about colors). While our color concepts are not of properties exactly like our color sensations, they are concepts of properties strongly tied to that phenomenal content. Hence, the subjectivity of our experiences is relevant when it comes to defining color properties.

27As I have been speaking of concepts of properties, let me stipulate this before going further: The colors are the properties to which our color concepts refer. If it so turns out that there are no properties corresponding to our color concepts, then our color concepts are misplaced or fail to refer to anything real. For a property to be the property of blueness, it must correspond well enough to our concept of blueness to be its proper referent. It does not have to be exactly like our current concept of blueness – we could be wrong about a thing or two – but presumably it must resemble it to a considerable degree. For this reason, I think we can say a great deal about properties such as colors on the basis of our concepts of them.

28We have some good reasons to think of colors as objective. According to color objectivism, objects have their colors independently of our color experiences. It is not a matter of our thoughts or perceptions whether and when a particular color is instantiated in an object. This seems consistent with many things we believe about colors. A flower growing wild that nobody has ever seen or shed any kind of thought can still be blue. Someone might wonder on what basis that flower is blue and providing the answer to that has been somewhat difficult for the color objectivists. That the flower is blue on the basis of its microphysical properties may seem dubious in light of the multiple realization of color and color metamerism. There is no single microphysical property (or even two or three… or two hundred) that the things we call blue have in common. However, the claim that the flower is blue, whatever the reason may be, still seems plausible.

  • 19  This is argued by Michael Tye 2006a. Tye claims that while we have no way of knowing which of us a (...)
  • 20  For an example of this, see Cohen, Hardin and McLaughlin 2006. This debate, started by Tye’s 2006a (...)

29There are also some good reasons for considering the colors subjective. One is that in some cases it seems to be a matter of opinion what color an object has. People disagree, for instance, about where exactly to put “true blue” or “true red” on the spectrum. Due to variation in human color vision, when asked to pick out a color chip that shows the true blue color, one person may select a chip that another considers slightly green or slightly purple. While some claim that this means that only some humans have correct color vision and thus correctly select the objectively true blue chip19, others take this to show that there is no such thing as true blue simpliciter20. This seems to suggest that at least in some cases, it is our perception – not which microphysical properties the object has – that determines whether a certain color is instantiated.

Subjective and Objective

30Determining whether colors are subjective or objective is no easy matter. One thing we may consider is whether the reason is simply that color terms are vague. That would make it possible for colors to be objective; the only issue would be that the boundaries where one color category ends and another begins were vague. I believe color terms are vague, but not that it solves the problem of true colors. A painted wall considered reddish orange by one person may be considered yellowish red by another. Their disagreement may be terminological; not over the color itself but over what to call it. But when one person perceives a color chip as truly blue while another perceives it as greenish blue, we have a case of differences in color perception. The two people have different experiences when they look at the color chip. They can still be in perfect agreement about the proper use of the term “blue”.

  • 21  An account along these lines can be found in Brown 2006, Maund 1995, and Rosenthal 1999.

31Another possible approach is to claim that color terms have multiple meanings. On such an account, the term “blue” has (at least) two possible referents, and it depends on the context which one is at play21. One referent is some kind of appearance property, the color the object in question looks to have, or blue-as-we-see-it. Another referent is an intrinsic property of the object, a spectral reflectance profile or even a specific microphysical property. So the term “blue” does not refer to just one property but to several properties of different types. This solves some problems. One is this: When I look at an object that I judge to be uniformly colored, say, a blue ball, there is also a sense in which it does not appear to have just one color. Parts of its surface may reflect light that is not blue; it may look white or green, for instance. So in one sense the ball looks to me as having various colors while in another sense it looks to me as just being blue. These multiple senses may occasionally invite misunderstandings between people but most of the time it is implied by the context of discussion which sense is at play and we can even switch back and forth between these senses without any problems.

32A multiple referent theory of color can also solve the disagreement I mentioned at the very beginning of the paper: people’s differing intuitions about imagined cases of radical changes in our color vision. On a multiple referent theory of color terms, it can be said that those who think the color of grass would have changed along with our color vision have one referent of “green” in mind whereas those who deny it have another referent in mind. In the first case, the referent of the term “green” is fixed by actual human experiences of green as they are now, whereas in the second, the referent of “green” is not fixed in this way.

33But a multiple referent theory does not solve the matter entirely. Compare the case of Jack and Jill looking at a color chip and disagreeing about whether it is truly blue or greenish blue with the case of their disagreeing about whether the chip is blue or red. In the first case both of them seem right. From the point of view of a multiple referent theory, this must be because we think Jack and Jill are speaking of the way the chip appears to them. The chip can have different appearance colors while it of course only has one set of spectral reflectances. But if so, why can we not say the same if Jack says the chip is blue while Jill says it is red? Why is there a problem when the chip looks blue to Jack and red to Jill if there is none when the chip looks truly blue to Jack and greenish blue to Jill? Why do we say that either Jack or Jill (or possibly both of them) must be misperceiving the color chip if one of them sees it as red while the other sees it as blue, but not if one of them sees it as truly blue while the other sees it as greenish blue?

34The judgments Jack and Jill are making depend on intrinsic features of the color chip. There is a certain range of intrinsic properties that allows for their bearers to be somewhere in the blue range. On the other hand, there is no range of intrinsic properties that makes it possible for one and the same thing to be blue or red, depending on the perceiver’s experiences. A certain set or range of intrinsic properties is the set of acceptable candidates for true blue. None of them is also a member of the set of acceptable candidates for red.

35Similar things can be said about some properties that tend to be considered subjective, such as funniness. When Jack says that Tina Fey is funny and Jill says she is not, we accept both judgments as valid because the properties Tina Fey has independently of Jack and Jill’s judgments are such as to make her an acceptable candidate for funniness. But suppose Jack says that losing a loved one is funny and Jill says it is not. These are not two equally valid judgments; Jack is plainly wrong. Losing a loved one is not an acceptable candidate for funniness. If Jack experiences amusement at losing a loved one, there is something wrong with his psychological state.

36An account of multiple referents is not the best way to deal with this, whether for colors, funniness, or any of the other properties I am sure we can all think of for which something similar holds. If we assume that the referents of color terms are fixed, we cannot explain how both Jack and Jill can be right when one of them says the chip is truly blue and the other says it is greenish blue. If we assume that the referents of color terms are not fixed, we cannot explain why something is wrong when Jack and Jill disagree over whether the chip is red or blue. To solve this, I propose that we think of the properties in question as both subjective and objective. Let me explain how:

37Our frequently mentioned color chip has various intrinsic features. Among them are features that put it in the blue color range and make it an acceptable candidate for true blue. The fact that these features are in place is mind-independent. It is not up to the thoughts or judgments of Jack or Jill, you or me that the chip has these features. However, it is mind-dependent whether these features make it so that the chip is truly blue or greenish blue. We might say that this is because the facts about that are perceiver-relative or we might say it is because a property such as true blue simpliciter cannot be instantiated; only true-blue-to-Jack, true-blue-to-Jill etc. Either way, this makes it subjective. But true blue is also objective to a considerable degree as the fact that the chip’s intrinsic features place it in the range of acceptable candidates for true blue is mind-independent.

38A similar case can be made for funniness. While it is a matter of opinion whether funniness is instantiated, we still differentiate between things that are potentially funny and things that are not. Losing a loved one, torture, and mass murder do not have mind-independent features that place them in the potentially funny category. Tina Fey, slipping on banana peel, and Lolcat pictures do.

39According to this account, there is an objectivity/subjectivity spectrum for properties. At one end, there are properties that are entirely objective, i.e. their instantiation is purely objective. At the other end there are properties that are entirely subjective. Properties that have a wide range of features for their acceptable candidates fall near the subjective end of the spectrum while those with a narrow range fall near the objective end of the spectrum. Consider a property such as being 75.3 cm long. There is only one way to fulfill the instantiation of that property. The perceiver of the object has no say when it comes to voting between hopeful candidates; there is only one who qualifies. This makes the instantiation of being 75.3 cm long entirely objective. True blue has a fairly narrow range of acceptable candidates. The subject has some say in which color chip is the one that is truly blue, but the restrictions determined independently of the perceiver are reasonably fixed. The range of acceptable candidates for funny things is wider; the subject has more choices than in the true blue case, and thus funniness gets a spot on the spectrum closer to the subjective end than blueness gets.

  • 22  It is possible that some might object to this idea of being both subjectively and objectively inst (...)

40This account does not exclude the possibility of color terms having multiple referents. That may very well be the case, and perhaps that is the best explanation of some apparent problems that arise when we speak of color. But multiple reference is not sufficient as a solution to all color issues. If we want an explanation of why disagreement about true blue vs. greenish blue is acceptable whereas disagreement about blue vs. red is not, thinking of colors as both subjectively and objectively determined is better22.

Possible differences between colors and other sensory properties

  • 23  Hopeful counterexamples include Smith 2007 and Nudds and O‘Callaghan Forthcoming.

41An examination of color is an examination of the sensory property that we access through vision. By sensory property I mean a property that we perceive with one of our senses, and associate with a certain sense modality. Our concept of it is to some degree based on the sensation we have when we perceive it. Colors are in this sense undoubtedly sensory properties. But vision is not our only mode of sensing the world. Odors, flavors, sounds, and heat are also sensory properties, associated with smell, taste, hearing, and touch. In recent decades, philosophical accounts of other sensory properties than colors have, however, been few and far between, although there have been some recent indications that the situation may be on the mend23.

42What I will consider is whether findings on color regarding its status as objective or subjective can be considered representative of all the sensory properties. Should we perhaps be dealing with each property separately? Can we assume that something analogous to the color account holds for each of the other sensory properties?

43While it would be unwise to extrapolate from everything there is to be said about color to claims about other sensory properties, I believe the differences there are between colors and other sensory properties do not affect the issue of whether the properties are objective or subjective.

44As I have explained, we have reasons for thinking of colors both as subjective and as objective. While our potentially differing color sensations play an important role in determining the colors of things, we expect color perception to be a means of access to the external world; to finding out about things that are not just figments of our imagination. This seems to hold for other sensory properties as well. Touching, smelling, tasting, and listening are all methods we use to find out about the way the world is. The sensations associated with them are also vital and we want the perceiving subjects to have a say when it comes to determining when the corresponding properties are instantiated. So as in the case of colors, there are some good reasons both for considering these properties objective and subjective. But before reaching a final verdict, we must examine some evidence to the contrary.

  • 24  I myself find this claim fairly plausible. Arguments for it can be found in Strawson P.F. 1959 and (...)

45Some have claimed that our sense modalities work in fundamentally different ways, and two such accounts will be described below. Sometimes, such accounts involve the idea that perception via some sense modalities is spatial, or yields conceptual materials we can use for forming spatial concepts, while perception via other modalities is non-spatial. Now, how would this affect the properties perceived by the sense modalities? This is how: Let us consider the Kantian claim that a conception of space is an essential element in our conception of mind-independence; of unperceived existence. Suppose this claim is true24. Then it is the case that the concept of an objective property, a property that can be instantiated independently of what we think or perceive, involves a spatial element. We must be able to make sense of the idea of the property being instantiated in an object that is located such that we cannot or do not perceive it; i.e. located “somewhere else”. The concept of a sensory property, however, is strongly tied to a certain sensory response associated with the appropriate modality. If the sense modality in question is non-spatial, it is unclear how the concept of a corresponding property could be spatial. Thereby, the concept – and thus the property – is missing an element essential to its objectivity. The conclusion is that properties perceived via non-spatial sense modalities must be exclusively subjective.

46If the argument I have just outlined is sound, properties perceived via non-spatial sense modalities cannot be objective. It is only via the spatial modalities that we can perceive properties that are objective, or to some extent objective. This suggests that we cannot infer from an account of color, a property perceived by vision, to properties perceived by other sense modalities. However, I do not think the argument is sound. Its weakness lies in the premise that there are fundamental differences between the sense modalities that make some of them spatial and some non-spatial.

  • 25  Strawson P.F. 1959: 169.

47In his book Individuals, P.F. Strawson claims that hearing is a non-spatial modality, whereas touch and vision are spatial. He creates an example of a purely auditory being, and argues that this being cannot possibly possess spatial concepts, the reason being that sounds are essentially non-spatial25.

48Evaluating the plausibility of Strawson’s proposal is difficult, especially because of difficulties imagining a purely auditory experience. A being who cannot see, smell, taste, or touch, and is unable to move – this is a being who cannot even feel her own body. Can trying to imagine the experiences of such a being really tell us much about what our actual auditory concepts are like? I find that doubtful. Before explaining why, I will consider another version of the claim that the sense modalities differ in this respect.

  • 26  Pasnau 1999: 313. I think we can safely assume here that Pasnau is referring to instances of the q (...)
  • 27  Pasnau 1999: 314.

49An account of sound considerably different from Strawson’s is proposed by Robert Pasnau. Pasnau claims that hearing is, like vision, a locational modality while touch and olfaction are not. Locational modalities, according to Pasnau, are those that «directly yield information not just about sensory qualities, but about the location of those qualities»26. On the other hand, says Pasnau; smell, taste, and touch are non-locational. He claims that we do not perceive smell as being located anywhere in particular and that the same thing holds for heat27.

50The difference between locational and non-locational sense modalities is, according to Pasnau, a fundamental one. Seeing and hearing yield information about the location of the properties seen or heard. Smelling, touching, and tasting do not yield such information; we do not perceive odor, heat and flavor as being anywhere in particular.

51If Pasnau is right, his view could lend support to an argument for the claim that colors and sounds have an objective element while other sensory properties do not. Pasnau himself does not make any such claim, but as discussed above, an argument of this kind seems salient if differences between sense modalities are presumed to make some of them spatial and some non-spatial. Being locational can be seen as a form of being spatial; therefore Pasnau’s locational modalities could fulfill the role of the spatial modalities while the non-locational modalities would be the non-spatial ones. What is interesting if we contrast Pasnau’s view with Strawson’s is, of course, that according to Strawson hearing is non-spatial and touch is spatial while on Pasnau’s account hearing is locational and touch non-locational.

  • 28  Pasnau 1999: 314.

52I find Pasnau’s account unconvincing from an empirical point of view. For instance, Pasnau claims that we do not perceive heat as existing at its source28. I cannot speak for others, but this most certainly counters my experiences of heat. Why Pasnau thinks there is some fundamental difference, locationally speaking, between an experience of heat and one of sound eludes me.

  • 29Ibidem.

53As support for his claims, Pasnau offers the example of the game hot/cold in which one searches for an object with the aid of clues in the form of hot and cold (hot means one is close to the object and cold that one is far from it). He describes the game as strictly inferential, as we do not use perception of heat to find the object. He then goes further and claims that this game «faithfully mirrors our perception of temperature, and also the way we perceive odours». In contrast, says Pasnau, when we search for the source of a sound, we do not move around randomly and infer, but use our hearing to perceive where the sound is coming from29.

54Again, Pasnau’s description is not consistent with my behavior or experiences. And surely, I am not the only one who has moved around randomly in order to find out if the high-frequency sound that is driving me nuts is coming from inside the house, from outside or inside my own ears. In addition, the claim that the game hot/cold is exactly like our actual experiences of heat and odor is unconvincing. The game is based on metaphoric use of the words “hot” and “cold”, and it is by no means clear why we should think of it as representative of actual perceptions of heat.

55Obviously, we do not smell or taste shapes (or hear them, for that matter), but our senses of smell and flavor can still give us information about location. We taste things as being inside our mouth or at least as touching our tongue. And although our sense of smell is not as acute as that of a dog, we can sometimes use it to locate things. Furthermore, there are good reasons to think that our spatial concepts are crossmodal or not tied to any particular modality.


  • 30  Menning et al. 2005; Kitagawa et al., 2005.
  • 31  Amedi et al. 2005; Kirchner and Colonius 2005.
  • 32  Macaluso and Driver, 2005; Meredith, 2002.
  • 33  Ghazanfar and Schroeder 2006.

56Recent research in neuropsychology indicates that we are better at locating things and at determining their shapes when more than one sense modality is involved. For instance, sound location becomes more acute when tactual cues are provided30. More generally, there seems to be vast evidence available for crossmodal integration, i.e. that our senses work better when they work together31, and that there are structural reasons for this in the neural system32. Research on brains of humans and other primates suggests that the neocortex is to a great extent multisensory33. This suggests that sensory integration takes place early in the perceptual process and thus that our experiences of the world are never unisensory but always to some extent crossmodal. The world is then not presented to us through one modality at a time, but through several of them jointly.

  • 34  For an account of the senses as different modes of awareness of the same environment, see Noë 2002

57These findings make any radical representational differences between our sense modalities unlikely. We do not seem to perceive the world piece by piece through the respective modalities and then put the pieces together. Instead, the evidence suggests that our perception of the world is a process relying on a joint effort of the different senses. Of course there are differences between the sense modalities, as is evident to us all. But the information we acquire does not seem to differ fundamentally from one sense to the other; this is where the senses work together34.

58Let us now return to the issue under consideration. The idea was that if some sense modalities are non-spatial while others are spatial, the properties perceived by the non-spatial modalities are non-spatial and thus missing an objective basis, and that this constitutes a difference between different sensory properties. But given the research findings cited above, there is no reason to think that any of our sense modalities are non-spatial. In this light, Strawson’s idea that we can imagine a being with only one sense modality and draw inferences on the basis of that makes little sense. Thus, neither Strawson nor Pasnau has provided any compelling reasons for making a distinction between spatial and non-spatial sense modalities.

A reason to make some kind of distinction?

59In his paper Phenomenal Character (Shoemaker 1994), Sydney Shoemaker compares bitterness with color and claims that colors are more objective than flavors:

  • 35  Shoemaker 1994: 32.

Consider Jonathan Bennett’s example of phenol-thio-uria, which tastes bitter to three-quarters of the population and is tasteless to the rest […] If as the result of selective breeding, or surgical tampering, it becomes tasteless to everyone, I say it has become tasteless. And if more drastic surgical tampering makes it taste sweet to everyone, I say it has become sweet. But I don’t think that if overnight massive surgery produces intrasubjective spectrum inversion in everyone, grass will have become red and daffodils will have become blue35.

  • 36  Shoemaker 1994: 37, fn. 6.

60The reason for this difference, according to Shoemaker, is that the semantics of color terms and flavor terms differ and that our color concepts are more objective than our flavor concepts. The idea seems to be that the reference of color terms is fixed by our actual color experiences whereas the reference of flavor terms is not. The semantics of flavor concepts is more strongly tied to their associated sensations than that of color concepts is36.

  • 37  My notion of property here is considerably broader than Shoemaker’s.

61I think this is an example where a multiple referent theory could be useful. Surely, we can say for each of the imagined cases that on one understanding of the term, the properties will have changed, and that on another understanding of it, they will not have changed37. There is a way to use “bitter” that makes it so that phenol-thio-uria will still be bitter, and there is also a way to use “bitter” that makes it so that phenol-thio-uria will no longer be bitter. Ditto for the color case. When and whether one understanding of the term is more valid than the other can be difficult to tell and on some occasions we may be uncertain which one is at play. Shoemaker’s intuition about the difference may suggest that we use the fixed versions of color terms more often than we do for flavor terms, but I very much doubt that we always use them in the case of color and never in the case of flavor.

62Another way to explain the difference is that taste properties may be nearer the subjectivity end of the spectrum than colors. This allows for elements of both subjectivity and objectivity in each.

63I have just argued that we do not have a reason to differentiate between our sense modalities when it comes to experiencing things in space or as having a location. Hence, if we were to consider, say, colors and sounds as something we perceive as having a location while odors, flavors and heat were perceived as being nowhere in particular (or even in ourselves), it would have to be on some other basis than an alleged difference between modalities. I do not believe there is such a basis.

  • 38  This paper is completed during my employment as a postdoctoral researcher at the Institute of Phil (...)

64What all sensory properties have in common is that the phenomenal content of our experiences of them plays a role in their definition. They are all properties we perceive objects around us as having, and that we think they have, at least some of the time and/or in some cases, independently of our perceptions of them. The exact circumstances of when and whether to think of the property as mind-dependent or mind-independent may vary to some degree from one property to another. But what all sensory properties share is sufficient to put them in the same category when we consider whether they are subjective or objective; i.e. the relationship of our senses to the world38.

Torna su


Blackburn, S.

– 1985, Error and the Phenomenology of Value, in T. Honderich (ed.), Morality and Objectivity: A Tribute to J.L. Mackie, London, Routledge and Kegan Paul: 1-22

Brown, D.H.

– 2006, On the Dual Referent Approach to Colour Theory, “Philosophical Quarterly”, 56: 96-113

Byrne, A. and Hilbert, D.R.

– 1997 Colors and Reflectances, in A. Byrne and D.R. Hilbert (eds.), Readings on color, Volume 1: The philosophy of color, Cambridge (Ma.), MIT Press: 263-288

– 2007, Truest blue, “Analysis”, 67: 87-92

Campbell, J.

– 1993, A Simple View of Colour, in J. Haldane and C. Wright (eds.), Reality, Representation, and Projection, Oxford, Oxford University Press: 257-268

Chalmers, D.

– 2006, Perception and the Fall from Eden, in T.S. Gendler and J. Hawthorne (eds.), Perceptual Experience, Oxford, Oxford University Press: 49-125

Cohen, J., Hardin, C.L. and McLaughlin, B.

– 2006, True Colours, “Analysis”, 66: 335-340

– 2007, The Truth about “The truth about true blue”, “Analysis”, 67: 162-166

De Anna, G.

– 2002, The Simple View of Colour and the Reference of Perceptual Terms, “Philosophy”, 77: 87-108

Edwards, J.

– 1998, The Simple Theory of Colour and the Transparency of Sense Experience, in C. Wright, B.C. Smith and C. MacDonald (eds.), Knowing Our Own Minds, Oxford, Oxford University Press: 371-389

– 2003, A Reply to De Anna on the Simple View of Colour, “Philosophy”, 78: 109-115

Evans, G.

– 1980, Things Without the Mind, in Z. van Straaten (ed.), Philosophical Subjects: Essays Presented to P.F. Strawson, Oxford, Oxford University Press: 76-116

Ghazanfar, A.A. and Schroeder, C.E.

– 2006, Is Neocortex Essentially Multisensory?, “Trends in Cognitive Sciences”, 10: 278-285

Johnston, M.

– 1992, How to Speak of the Colors, “Philosophical Studies”, 68: 221-263

Kitagawa, N., Zampini and M., Spence, C.

– 2005, Audiotactile Interactions in Near and Far Space, “Experimental Brain Research”, 166: 528-537

Macaluso, E. and Driver, J.

– 2005, Multisensory Spatial Interactions: a Window onto Functional Integration in the Human Brain, “Trends in Neurosciences”, 28: 264-271

Maund, B.

– 1995, Colours: Their Nature and Representation, Cambridge, Cambridge University Press

McGinn, C.

– 1996, Another Look at Color, “Journal of Philosophy”, 93: 537-553

Menning, H. et al.

– 2005, Spatial Auditory Attention is Modulated by Tactile Priming, “Experimental Brain Research”, 164: 41-47

Meredith, M.A.

– 2002, On the Neuronal Basis for Multisensory Convergence: a Brief Overview, “Cognitive Brain Research”, 14: 31-40

Noë, A.

– 2002, On What We See, “Pacific Philosophical Quarterly”, 83: 57-80

Nudds, M. and O’Callaghan, C. (eds.)

– Forthcoming, Sounds and Perception: New Philosophical Essays, Oxford, Oxford University Press

Pasnau, R.

– 1999, What is Sound?, “Philosophical Quarterly”, 49: 309-324

Rosenthal, D.

– 1999, The Colors and Shapes of Visual Experiences, in D. Fisette (ed.), Consciousness and Intentionality: Models and Modalities of Attribution, Western Ontario Series in the Philosophy of Science 92, Dordrecht, Kluwer: 95-118

Shoemaker, S.

– 1994, Phenomenal Character, “Noûs”, 28: 21-38

Smith, B.C. (ed.)

– 2007, Questions of Taste: The Philosophy of Wine, Oxford, Oxford University Press

Strawson, G.

– 1989, Red and “Red”, “Synthese”, 78: 193-232

Strawson, P.F.

– 1959, Individuals. An Essay in Descriptive Metaphysics, London, Methuen

Stroud, B.

– 2000, The Quest for Reality: Subjectivism and the Metaphysics of Colour, Oxford, Oxford University Press

Triplett, T.

– 2007, Tye’s Missing Shade of Blue, “Analysis”, 67: 166-170

Tye, M.

– 2006a, The Puzzle of True Blue, “Analysis”, 66: 173-178

– 2006b, The Truth about True Blue, “Analysis”, 66: 340-344

Torna su


1  We can suppose that the virus causes genetic changes as well so that future generations of humans will be born with the same kind of visual system as their ancestors who got the virus.

2  See Blackburn 1985: 14 for someone whose intuition seems to be that the color of grass would have changed. The intuition of Shoemaker 1994: 32 is quite the opposite.

3  An example of such a view can be found in Byrne and Hilbert 1997.

4  Johnston 1992: 138.

5  Examples of color primitivists are Campbell 1993, McGinn 1996, Strawson G. 1989, Stroud 2000, and De Anna 2002.

6  Error theorists include Boghossian and Velleman 1989, Mackie 1976, and Hardin 1988.

7  Campbell 1993: 257.

8  Campbell 1993: 258.

9  McGinn 1996: 538.

10  Edwards 1998; 2003. Arguments against Edwards 1998, to which Edwards responds in his 2003, can be found in De Anna 2002.

11  Edwards 2003: 110.

12  A related criticism can be found in Chalmers 2006: 67-68.

13  Chalmers 2006: 50.


15  Chalmers 2006: 66-69.

16  Chalmers 2006: 72.

17  Of course, not just any causal chain will do in this respect. The father could for instance cause the child to draw a picture representing something completely different. I take it that in addition to the causal chain some degree of resemblance of the picture to the father may be needed, as well as an intent on the child’s behalf to have the picture represent his father.

18  My view here is similar to that of Johnston’s mentioned above.

19  This is argued by Michael Tye 2006a. Tye claims that while we have no way of knowing which of us are the lucky right perceivers of true blue, the fact of the matter is that some of us are.

20  For an example of this, see Cohen, Hardin and McLaughlin 2006. This debate, started by Tye’s 2006a, is continued in Tye 2006b, Byrne and Hilbert 2007, Cohen et. al 2007, and Triplett 2007.

21  An account along these lines can be found in Brown 2006, Maund 1995, and Rosenthal 1999.

22  It is possible that some might object to this idea of being both subjectively and objectively instantiated and prefer to speak of restricted subjectivity instead. I am not certain whether such an objection would mainly be terminological or run deeper than that.

23  Hopeful counterexamples include Smith 2007 and Nudds and O‘Callaghan Forthcoming.

24  I myself find this claim fairly plausible. Arguments for it can be found in Strawson P.F. 1959 and Evans 1980.

25  Strawson P.F. 1959: 169.

26  Pasnau 1999: 313. I think we can safely assume here that Pasnau is referring to instances of the qualities being located somewhere. In the following discussion, this will be the implied use.

27  Pasnau 1999: 314.

28  Pasnau 1999: 314.


30  Menning et al. 2005; Kitagawa et al., 2005.

31  Amedi et al. 2005; Kirchner and Colonius 2005.

32  Macaluso and Driver, 2005; Meredith, 2002.

33  Ghazanfar and Schroeder 2006.

34  For an account of the senses as different modes of awareness of the same environment, see Noë 2002.

35  Shoemaker 1994: 32.

36  Shoemaker 1994: 37, fn. 6.

37  My notion of property here is considerably broader than Shoemaker’s.

38  This paper is completed during my employment as a postdoctoral researcher at the Institute of Philosophy at the University of Iceland. For comments on earlier versions of it, I am grateful to Matti Eklund, Emily Esch, Tamar Szabó Gendler, and Nicholas Sturgeon.

Torna su

Per citare questo articolo

Notizia bibliografica

Eyja M. Brynjarsdóttir, «Stuck in the middle: Colors between the subjective and the objective»Rivista di estetica, 43 | 2010, 47-65.

Notizia bibliografica digitale

Eyja M. Brynjarsdóttir, «Stuck in the middle: Colors between the subjective and the objective»Rivista di estetica [Online], 43 | 2010, online dal 30 novembre 2015, consultato il 24 juin 2024. URL:; DOI:

Torna su

Diritti d’autore


Solamente il testo è utilizzabile con licenza CC BY-NC-ND 4.0. Salvo diversa indicazione, per tutti agli altri elementi (illustrazioni, allegati importati) la copia non è autorizzata ("Tutti i diritti riservati").

Torna su
Cerca su OpenEdition Search

Sarai reindirizzato su OpenEdition Search